Study Finds Manipulation Signals in Bitcoin Markets

United States.- 17 July 2026 www.zonadeazar.com Research from Stanford University and Singapore Management University has identified trading patterns consistent with settlement-price manipulation in Polymarket’s five-minute Bitcoin prediction markets.

The academic study, written by David Dai, Ruizhe Jia and Shihao Yu, examined approximately 16,000 Bitcoin contracts during the two months following the launch of the five-minute markets on 12 February 2026.

The contracts allow participants to predict whether Bitcoin’s price will finish above or below the level recorded at the beginning of each period.

Trading activity before settlement

The researchers detected concentrated increases in Bitcoin buy and sell orders on Binance during the final seconds before contracts expired.

These trades temporarily moved the price used to settle the markets, before the movements rapidly reversed following settlement.

In contracts whose outcome remained close to evenly balanced shortly before expiry, order flow during the final seconds was up to 3.9 times greater than during normal settlement periods.

Estimated profits of US$8.2 million

The authors identified 821 accounts that generated profits during cycles classified as likely to contain manipulation, from approximately 243,000 participants.

According to the study, these accounts produced combined estimated profits of US$8.2 million, while retail users reportedly absorbed 93% of the corresponding losses.

The researchers note that their identification is based on transaction and profit patterns rather than direct evidence of each trader’s intent.

Vulnerability in short-term contracts

The potential vulnerability arises because the contracts settle using a Chainlink reference price compiled from data across several exchanges.

A participant holding a significant position could trade Bitcoin in the spot market during the final seconds in an attempt to move the settlement price in their favour.

The same patterns were substantially weaker in 15-minute contracts. The researchers therefore propose longer market horizons and the use of time-weighted average prices instead of a single price recorded at settlement.

The paper, entitled Settlement Manipulation in Prediction Markets, remains an academic working paper and its findings will require further analysis.

Edited by: @_fonta